Abstract
This study assessed treatment change patterns in Parkinson’s disease psychosis (PDP) residents receiving antipsychotic (AP) therapies in U.S. long-term care (LTC) facilities. Residents with PDP in LTC between 01/01/13 and 06/30/16 were identified with ≥1 claim of psychosis, hallucinations, or delusions after PD diagnosis. Treatment patterns were evaluated during the 12 months post index. We identified 864 PDP residents: 408 (47.2%) on AP therapy and 456 (52.8%) on no AP therapy. A total of 335 residents (82.1%) continued, 13 (3.2%) discontinued, 11 (2.7%) switched, and 49 (12.0%) augmented (used ≥2 APs) their index AP therapy. Based on the multivariate regression analysis, younger age, male gender, anemia, anxiolytic use or anxiety, sedatives/hypnotic use, bladder disorders including urinary tract infections, coronary conditions, diabetes, hypertension, and dementia were associated with a higher likelihood of treatment change. Understanding the factors associated with treatment change may inform ways to improve management of PDP in the U.S. LTC setting.
Introduction
Parkinson’s disease (PD) is a progressive neurodegenerative disorder characterized by the loss of nigral dopaminergic neurons and is associated with rigidity, resting tremor, and bradykinesia (Jakel, 2014). A recent study released by the Parkinson’s Foundation found that the prevalence of PD is increasing, estimating that nearly 1 million Americans over age 45 will be diagnosed with PD by 2020, a number expected to rise to 1.24 million by 2030 (Marras et al., 2018). The study found that men are at higher risk than women and that risk for everyone increases with age (Marras et al., 2018). An estimated 40% of these patients have Parkinson’s disease psychosis (PDP), which is characterized by hallucinations and delusions, is associated with significant caregiver burden, morbidity and mortality to the patient, and is a major reason for nursing home (NH) placement among Parkinson’s patients (Aarsland et al., 2000; Goetz & Stebbins, 1993; Jethwa & Onalaja, 2015). Psychosis defined with delusions and/or hallucinations is a common nonmotor feature of PD and can be a reflection of the primary progression of the underlying disease process (Aarsland et al., 2000). First-generation and second-generation antipsychotic (AP) therapies are frequently used off-label for the treatment of PDP, including quetiapine, risperidone, olanzapine, haloperidol, and clozapine, the latter of which is associated with burdensome monitoring (Seppi et al., 2011). Off-label AP use is common in patients with PDP; however, APs can worsen Parkinsonism and cause sedation, orthostasis, metabolic changes, weight gain, dry mouth, or memory impairment (Combs & Cox, 2017; Sarva & Henchcliffe, 2016); in addition, the evidence of efficacy is limited (Goetz & Stebbins, 1993; Seppi et al., 2011; Weintraub et al., 2011).
Treatment options have been somewhat limited in patients who develop PDP (Yuan et al., 2017). It has been recommended that before initiating treatment with atypical APs to manage the symptoms of PDP, reduce and, where possible, discontinue medications that can worsen psychosis (Ravina et al., 2007; Sarva & Henchcliffe, 2016; Yuan et al., 2017; Zahodne & Fernandez, 2008). Currently, there is only one treatment, pimavanserin, approved by the U.S. Food and Drug Administration (FDA) in April 2016 for hallucinations and delusions associated with PDP. Pimavanserin preferentially targets the 5-HT2A receptors, which are thought to play an important role in PDP (Combs & Cox, 2017; Sarva & Henchcliffe, 2016). The unique pharmacology of pimavanserin establishes a new drug class—selective serotonin inverse agonists (SSIA)—by not only preferentially targeting 5-HT2A receptors but also avoiding activity at dopamine and other receptors commonly targeted by APs (Combs & Cox, 2017; Sarva & Henchcliffe, 2016). Typical PD therapy consists of drugs that stimulate dopamine to treat patients’ motor symptoms, such as tremor, muscle rigidity, and difficulty with walking. Pimavanserin does not interfere with patients’ dopaminergic therapy, and therefore does not impair their motor function, and has other advantages, including no sedation or orthostasis, metabolic changes, weight gain, dry mouth, or memory impairment (Combs & Cox, 2017; Sarva & Henchcliffe, 2016).
There are limited studies evaluating treatment of PDP especially in the long-term care (LTC) or nursing home (NH) setting. The current off-label options (clozapine, quetiapine, risperidone, and olanzapine) are not FDA approved or indicated for PDP, and furthermore, they all have the propensity to exacerbate PD motor symptoms (Bloomfield et al., 2012; Chou et al., 2007; Shotbolt et al., 2010; Yuan et al., 2017). In addition, there are limited data in the literature on evaluating the treatment patterns and assessing factors associated with treatment change within a PDP population initiated on APs (PDP-AP) residing in the NH setting. In this retrospective study, we identified patients with PDP prescribed APs and aimed to (a) describe their clinical characteristics and the type of AP therapies prescribed, (b) assess the treatment patterns, and (c) evaluate the factors associated with treatment change in patients with PDP who were on AP treatment.
Materials and Methods
Study Design and Data Source
A retrospective cohort study was conducted using a national U.S. LTC pharmacy database consisting of NH residents with PD. The database was linked to an E-Census minimum data set, which captures diagnosis data through electronic health records. Patients with PD and PDP were identified between January 1, 2013, to June 30, 2016. Once the PDP population was identified, the first AP prescription was labeled in the pharmacy database and defined as index date. The study start date was January 1, 2012, to include 12 months of NH enrollment and to identify baseline characteristics during the 12 months preindex; the end date of follow-up was June 30, 2017, which evaluated the study outcomes during the 12 months postindex. This study did not involve individually identifiable data; institutional review board approval and patient consent was not required.
Study Population
Initially, patients with PD were identified during January 1, 2013, to June 30, 2016, with ≥2 diagnosis codes (332.xx, 332.0x, G20.xx) at least 30 days apart. The first PD diagnosis was labeled as the PD date. To avoid including patients receiving an AP for conditions other than PDP, patients were excluded if they had any history of secondary parkinsonism, dementia with Lewy bodies, primary psychiatric disorders (schizophrenia, schizoaffective disorder), other mood disorders with psychotic features, or delirium at any point in the study time period (January 1, 2013, to June 30, 2016); additionally, patients with delusional disorder on or before the PD date were excluded. Next, PDP patients with ≥1 diagnosis of psychosis, hallucinations, or delusions after the PD date were identified. Using the pharmacy prescription claims data, the primary AP prescribed on or after the PDP date was identified, and all pharmacy data within the 12 months post index were extracted.
Study Outcomes
Treatment change
A treatment change during the 12 months postindex was evaluated, starting with the primary index AP. All patients with an AP prescription were followed until a change in their primary AP therapy occurred (switch, discontinue, or augmentation), or they were followed until the end of their respective 12 months as continuation. Patients who continued their index AP therapy were defined based on a 30-day gap (end date of last index fill + days of supply + 30 days). If a patient continued index therapy within the 30-day gap, they were labeled as a continuer. Patients who switched their index AP therapy within the 30-day gap were defined as starting another AP therapy and labeled as a switcher. Patients who did not start any AP therapy in the 30-day gap were labeled as discontinuers. Finally, any patient who had an addition of an AP therapy was labeled as augmenter. All patients had two prescriptions of the same or different APs to ensure they were on continuous APs.
Mean time in days for treatment change
Once the treatment change was determined, the mean time to treatment change was calculated. Each prescription had a start date of the primary AP, and an end date was calculated as the last prescription date + days of supply.
Statistical Analysis
Two study groups were created: PDP AP therapy and PDP No-AP therapy. Descriptive statistics (N, mean, SD, percentages) were used to compare the baseline characteristics between the two study groups; PDP AP therapy patients were then categorized into continuers, switchers, discontinuers, and augmentation groups. Categorical variables were analyzed using chi-square tests, and differences in mean values of continuous variables were analyzed using Student’s t test. A backward selection, multivariable, logistic regression model was used to identify factors associated with treatment change (switchers + discontinuers + augmenters) versus patients who continued their AP therapy. The multivariate logistic regression model controlled for characteristics such as age, sex, comorbid conditions, and concomitant medications. The time to a treatment change was evaluated using descriptive statistics, such as mean days, SD. A p value of <.05 was considered statistically significant, and the SAS 9.4 software version was used for all analyses (SAS Institute, Cary, NC).
Results
Study Population
In total, 7,440 residents with PD were identified. After exclusion criteria were applied, 6,241 residents remained in the study cohort, of whom 864 residents (13.8%) had PDP (Figure 1). There were 408 (47.2%) residents with PDP on AP therapy (PDP-AP) and 456 (52.8%) residents with PDP not on AP therapy (PDP No-AP). The PDP-AP group was younger (76 years) versus the PDP No-AP group (78 years), p < .001 (Table 1). Comorbidities included in the model of treatment change are shown in Table 1. Among the PDP-AP group, the most common comorbidities were hypertension (74%), dementia (54%), falls (53%), and anxiety (53%). The most used concomitant medications were acetylcholinesterase inhibitors (64%), anxiolytics (65%), and antidepressant agents (52%) (Table 1). The prevalence rates of anxiety, bladder disorders (including urinary tract infections), dementia, depression, diabetes, hypertension, renal disease, and stroke were higher (p < .05) within the PDP-AP group versus PDP No-AP group (Table 1). Usage rates of anxiolytics, antidepressants, sedatives/hypnotics, and antidementia medications were higher within the PDP-AP group (p < .05) (Table 1).
Patient study cohort diagram.
Baseline Characteristics of PDP Patients on AP and Not on AP.
Note. PDP = Parkinson’s disease psychosis; AP = antipsychotic; SD = standard deviation; UTI = urinary tract infection; COPD = chronic obstructive pulmonary disease.
p value < .05 considered statistically significant.
Index AP Therapies
The most common index AP therapies and mean daily dose (MDD) were quetiapine (52%, MDD 75 ± 65 mg), risperidone (17%, MDD 9 ± 3 mg), olanzapine (11%, MDD 10 ± 5 mg), aripiprazole (9%, MDD 10 ± 9 mg), and haloperidol (6%, MDD 14 ± 10 mg) (Table 2). Within the quetiapine group, the MDD for the majority of residents (187 residents, 88%) was <100 mg.
Mean Daily Dose for Primary AP.
Note. AP = antipsychotic; PDP = Parkinson’s disease psychosis; SD = standard deviation.
Augmentative therapies are included in their respective categories; patients are not mutually exclusive since some patients are on two therapies as primary AP.
Study Outcomes
Treatment patterns and mean days to treatment change
There were 335 patients (82%) found to continue on their index AP therapy, and 73 patients (18%) had a treatment change (11 patients switched, 49 patients augmented an AP, and 13 patients discontinued). Residents who continued their index AP were on their therapy for a mean (SD) of 178 (138.3) days. Residents who switched their index AP did so after a mean (SD) of 141 (109.6) days. For those discontinuing or augmenting their therapies, residents discontinued their index AP therapy after a mean (SD) of 69.8 (59.3) days or augmented their index AP therapy after a mean (SD) of 31.2 (24.1) days.
Factors associated with treatment change
PDP-AP patients who had a treatment change were younger in age and had more comorbidities and a higher percentage of concomitant medications (Table 3). Based on the multivariate regression analysis, the following factors were associated with higher likelihood of treatment change: younger age, male sex, anemia, sedative/hypnotic use, anxiolytic use or anxiety, bladder disorders (including urinary tract infections), coronary conditions, diabetes, hypertension, and dementia (Table 4).
Baseline Characteristics of PDP Patients Continuing on Index AP Versus Treatment Change.
Note. PDP = Parkinson’s disease psychosis; AP = antipsychotic; SD = standard deviation; UTI = urinary tract infection; COPD = chronic obstructive pulmonary disease.
Treatment change defined as switch, discontinue, or augmentation.
p value < .05 considered statistically significant.
Multivariate Logistic Regression of Factors Associated With Treatment Change.
Note. UTI = urinary tract infection; ACOi = acetylcholinesterase inhibitors.
Indicates which factors are found to be significant.
Discussion
This study assessed PDP patients with or without AP therapies in U.S. LTC skilled nursing facilities and evaluated the factors associated with treatment change. To our knowledge, this is the first study evaluating PDP disease, the use of off-label APs, concomitant comorbidities, and medications for NH residents with PDP within a national LTC database.
It is important to discuss prior and current criteria or guidelines and provide more context to our findings (American Geriatrics Society Beers Criteria Update Expert Panel, 2015, 2019; Seppi et al., 2019). The time period for the data evaluated in this study is from 2013 to 2017. During this time, the American Geriatrics Society (AGS) 2015 Beers Criteria® recommended that “all APs (except aripiprazole, quetiapine, and clozapine) were to be avoided with the rationale that dopamine-receptor antagonists had the potential to worsen parkinsonian symptoms” (American Geriatrics Society Beers Criteria Update Expert Panel, 2015); in addition, “quetiapine, aripiprazole, and clozapine appeared to be less likely in precipitating worsening of PD” (American Geriatrics Society Beers Criteria Update Expert Panel, 2015). AGS recently updated the Beers Criteria (American Geriatrics Society Beers Criteria Update Expert Panel, 2019), discussing the evidence on APs to treat psychosis in PD patients, and the panel decided to remove aripiprazole and add pimavanserin. Currently, the 2019 AGS Beers Criteria “recognize quetiapine, clozapine, and pimavanserin as exceptions to the general recommendation to avoid all APs in older adults with PD” (American Geriatrics Society Beers Criteria Update Expert Panel, 2019). Furthermore, the 2019 AGS Beers Criteria state that pimavanserin and clozapine appear to be less likely to precipitate worsening of PD; however, “quetiapine has only been studied in low-quality clinical trials with efficacy comparable to that of placebo in five trials and to that of clozapine in two others” (American Geriatrics Society Beers Criteria Update Expert Panel, 2019).
The Movement Disorder Society (MDS)-commissioned evidence-based medicine review for the treatments of PDP has also been updated (Seppi et al., 2019). Three new studies were evaluated, and recommendations were discussed because practice implications have been changed since the previous review, especially related to quetiapine. The update stated that “there is insufficient evidence for quetiapine to be rated for the treatment of psychosis in PD” and “there are no high-quality randomized controlled trials (RCTs) available for quetiapine for the treatment of psychosis in PD, and quetiapine was similarly efficacious to clozapine in a clozapine-controlled trial that did not include a placebo arm” (Seppi et al., 2019). Olanzapine was evaluated in a low-quality study with negative results, and as such the conclusions were “non-efficacious” and “not useful” (Seppi et al., 2019).
In this study, fewer than half (47.2%) of the PDP residents were treated with any APs, and the majority of prescribed APs were not aligned with the 2019 AGS Beers Criteria and MDS Commissioned Review guidelines (American Geriatrics Society Beers Criteria Update Expert Panel, 2019; Seppi et al., 2019). In addition, we saw that more than 50% of the residents with PDP remained untreated with any APs, placing them at risk for poor symptom control and outcomes. Patients in the PDP-AP group were younger in age, had higher comorbidities, and a higher use of concomitant medications relative to the PDP No-AP group. Further research is needed to understand this patient population and to compare our findings with other data sources.
The number of patients not receiving AP treatments might be explained in part by guidelines from the Centers for Medicare and Medicaid Services, which has focus on reducing off-label AP use in older patients (Centers for Medicare & Medicaid Services, 2016). Medications being used in this population must be reviewed frequently, and attempts to taper the dosage should be made for patients receiving APs to ensure that the patient is receiving the appropriate dose and to evaluate whether the medication could be withdrawn. While this practice has generally reduced overall AP use in U.S. LTC (Maust et al., 2018), it may have led to changes in prescribing practices of alternative medications and an increase in the number of patients not receiving treatment for their psychosis.
The top five APs identified in this study were quetiapine (52%), risperidone (17%), olanzapine (11%), aripiprazole (9%), and haloperidol (6%). Quetiapine was the most prescribed and is approved for the treatment of schizophrenia, bipolar disorder, and as adjunct therapy for major depressive disorder. Five RCTs were conducted to establish the efficacy of quetiapine in PDP (Fernandez et al., 2009; Kurlan et al., 2007; Ondo et al., 2005; Rabey et al., 2007; Seppi et al., 2019; Shotbolt et al., 2009, 2010). Four of the RCTs found no evidence of quetiapine efficacy in patients with PDP, and the only positive study included just 16 patients and excluded patients with delusions, which are more difficult to control (Fernandez et al., 2009). In addition, two single-blind randomized trials compared quetiapine and clozapine and showed a significant improvement of 25% to 30% in both treatment groups as assessed by the Brief Psychiatric Rating Scale and Clinical Global Impression Scales (Jethwa & Onalaja, 2015; Seppi et al., 2019; Shotbolt et al., 2010). The results of these studies suggest that clozapine and quetiapine have similar efficacy; however, according to the newly updated MDS Commissioned Review (Seppi et al., 2019), clozapine is regarded as clinically useful but needing specialized monitoring, which can be burdensome to the patients and caregivers, and quetiapine was determined to have insufficient evidence for efficacy and to be only possibly useful (Jethwa & Onalaja, 2015; Seppi et al., 2019). In the study reported here, the use of clozapine was very low, perhaps due to the monitoring burden.
The mean daily dose for quetiapine was 74.5 mg, with more than half of the residents prescribed ≤25 mg, a dose that could be used to aid patients with sleep disturbances (Carr et al., 2016). Prior studies have evaluated the use of low-dose quetiapine and its risk on metabolic consequences as well (Carr et al., 2016; Shotbolt et al., 2010). The findings of these studies discuss that there may be negative metabolic consequences (increase in blood pressure, weight, glucose levels), and prescribing low-dose quetiapine for sleep as first-line medication should be avoided (Carr et al., 2016; Shotbolt et al., 2010). Weintraub et al. (2011) also found quetiapine to be the most prescribed AP for PDP patients. Overall, NH residents with PDP are receiving nonoptimal AP therapies that are not supported by consensus guidelines.
Approximately 80% of residents continued on their index therapy and did not have any treatment change. This finding was similar to Weintraub et al. (2011) where the majority of patients with PDP had no change in their AP during their observation period. The residents who had treatment change were younger in age and had more comorbidities and higher use of concomitant medications. For example, we found these residents to be male, anemic, to have congestive heart disease and diabetes, and to be taking anxiolytics. Further research is needed to evaluate or compare this finding with data from other databases. There were 54% PDP-AP patients with a diagnosis of dementia. Dementia is increasingly recognized as a common long-term complication of PD, occurring in 80% to 90% of patients (Aarsland et al., 2003; Weintraub et al., 2011). This finding is a concern, given the association between AP use and increased morbidity and mortality in patients with dementia (Maust et al., 2015; Ralph & Espinet, 2018; Schneider et al., 2005).
Every database research study has some limitations to address. Due to the nature of the database, there could be coding or data input errors. The number of PDP patients (13.8%) seemed very low when compared to the prevalence in prior studies, and the assumptions for this could be that patients could be underdiagnosed for psychosis or not being treated. It is possible that they could be treated by nonpharmacological therapies. There were approximately 11% of non-PDP patients on APs, and these patients could potentially be PDP patients, increasing our sample to approximately 24%; however, we could not identify a psychosis diagnosis. We also do not know how long patients had a history of PD, thus this variable could not be added as an independent variable and controlled for in the statistical model. We did not have access to chart notes to determine reasons for treatment change. Not all LTC facilities had E-census data, and we could not evaluate all the patients in their service; therefore, further studies will be needed using other databases, such as closed-integrated databases, to compare treatment change rates and continuation rates of APs with those found in this study.
Conclusion
A high percentage of off-label APs were prescribed in the LTC setting to treat PDP despite equivocal efficacy and questionable safety, especially haloperidol, a first-generation AP (Chou et al., 2007; Seppi et al., 2019). These data highlight the need for increased awareness of appropriate AP selection to support optimal treatment of NH residents with PDP. In this study, PDP patients were treated with off-label APs, and, according to the MDS-commissioned evidence-based update on treatments of PDP (American Geriatrics Society Beers Criteria Update Expert Panel, 2019), quetiapine was found to have insufficient evidence for efficacy, olanzapine was found to be not efficacious, and risperidone was not evaluated because it did not include the MDS requirement (i.e., studies in PDP had to have a minimum of 20 patients who were treated for a minimum of 4 weeks). Currently, pimavanserin is the only U.S. FDA-approved treatment for hallucinations and delusions associated with PDP and is noted as such in the MDS-commissioned evidence-based update on treatments of PDP (American Geriatrics Society Beers Criteria Update Expert Panel, 2019). Further research is needed to optimize PDP treatment and outcomes within the LTC setting.
Footnotes
Acknowledgements
The authors acknowledge the helpful study insights from Kurt Grady, PharmD (prior employee of Acadia Pharmaceuticals). They also acknowledge Kayla Mills, HEOR pharmacist intern of Pharmacore Pharmacy Services, for her support in performing the literature review and data review and assistance with development of the manuscript. Acadia Pharmaceuticals Inc. funded this analysis and participated in the study design and analysis and reviewed the manuscript.
Authors’ Note
Andrew Shim’s current affiliation is Arena Pharmaceuticals, San Diego, CA, USA.
Author Contributions
Study concept and design: N.R., A.S., S.A., and V.A. Acquisition of data: S.Q. Analysis of data: N.R. Interpretation of data: N.R., A.S., S.A., S.Q., and V.A. Drafting of the manuscript: N.R. Critical revision of the manuscript for important intellectual content: N.R., A.S., S.A., S.Q., and V.A.
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: N.R., S.A., and V.A. are employees of Acadia Pharmaceuticals Inc. A.S. was an employee of Acadia Pharmaceuticals Inc. at the time of the analysis. S.Q. was Chief Clinical Officer at PharMerica at the time of these analyses; she has been an unpaid consultant to Acadia Pharmaceuticals Inc.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Acadia Pharmaceuticals Inc.
Data Sharing Statement
Data available on request from authors: The data that support the findings of this study are available from the corresponding author upon reasonable request.